Chromatography Forum

Hello all,

As the title says, I've been experiencing two separate issues pertinent to the manual injection method I'm using for analyzing my samples:

1. The carryover peaks recorded for blank injections is larger than the peaks recorded for the actual sample

2. Adding a "wash" injection to mitigate problem (1) dilutes the peaks recorded for the sample

I've narrowed down the source of problem to the glass syringe and/or Rheodyne 7725i sample injector but I'm unable to solve this problem which has resulted in excessive delay in my work. Does anyone have any idea regarding:

A. What is causing the carryover peaks to be larger than the sample peaks?

B. How can I solve this problem with the carryover peaks without replacing it with problem (2)?

For information, please find the pertinent information regarding the situation below:

1. Sample: Monosaccharide fermentation broth incl. organic acids & alcohols
2. HPLC: Shimadzu Prominence series
3. Column: RHM column (Pheonomenex)
4. Mobile Phase: H2SO4 (0.03N)
5. Column temperature: 40 degrees
6. Detector temperature: 40 degrees
7. Flow Rate: 0.6 mL/min
8. "Binary" gradient pumping same mobile phase through both pumps*
9. Sample loop volume: 20 uL
10. Injection volume: 150 uL

Notes:
* The "binary" gradient is being used to mitigate the problems with frequent pump failure caused by air bubbles
- The null injection does not produce any peaks, so I've ruled out the column as a source of the problem.
- I do wash the glass syringe with water and/or the mobile phase before every injection, & occasionally the sample injector too.
- Not sure if this issue is due to azeotrophy of the compounds with water and/or mobile phase
- Problem mostly observed with glucose and lactic acid.
- I've observed that lack of liquid draining out from the injector upon overfilling the sample loop

Some comments and questions to think about:

Q: Why are you injecting 150 ul into a 20 ul loop ? The accepted method of 3x loop volume for a full volume injection should be followed (in this example, 60ul total). Just odd to use 150ul for no reason ...

You describe "carryover", but I do not see any mention that you serviced whatever type of injection valve you are using. Valves require regular maintenance. Failure to maintain the value results in carryover. If you have not done so already, please properly service the valve (always a good idea to have the spare parts on-hand as this is a routine service).

Have you tried a true blank injection as part of you troubleshooting (a "Blank" injection involves injecting the mobile phase in the same volume as your sample)? The results are often very diagnostic.

From what you describe, either you are using the injector in an unusual way (plumbed wrong, wrong parts, wrong positions) OR far more likely the injector is worn out/broken.

Column fouling may also be contributing to the problems too.

"Wash" injections (you wrote: "dilutes the peaks recorded for the sample") , that is a follow up injection of mobile phase should never "dilute" any samples. I have no idea what you are referring to in this context. IN fact, for most types of analysis, a needle wash should not be needed unless the samples are rather 'sticky'.

You wrote: "The "binary" gradient is being used to mitigate the problems with frequent pump failure caused by air bubbles"; What gradient? It appears you are running isocratic with a RID (?). Pump cavitation is often caused by failure to maintain prime of the system. Common reasons for this include: a lack of continuous degassing of the mobile phase; sticking inlet or outlet valves; piston seal failure(s); clogged pickup frits; leaks or loose fittings. *If you are running a separate gradient step to wash the column down, what solvents are you using coupled to what time/flow program?

Some comments and questions to think about:

Q: Why are you injecting 150 ul into a 20 ul loop ? The accepted method of 3x loop volume for a full volume injection should be followed (in this example, 60ul total). Just odd to use 150ul for no reason ...

You describe "carryover", but I do not see any mention that you serviced whatever type of injection valve you are using. Valves require regular maintenance. Failure to maintain the value results in carryover. If you have not done so already, please properly service the valve (always a good idea to have the spare parts on-hand as this is a routine service).

Have you tried a true blank injection as part of you troubleshooting (a "Blank" injection involves injecting the mobile phase in the same volume as your sample)? The results are often very diagnostic.

From what you describe, either you are using the injector in an unusual way (plumbed wrong, wrong parts, wrong positions) OR far more likely the injector is worn out/broken.

Column fouling may also be contributing to the problems too.

"Wash" injections (you wrote: "dilutes the peaks recorded for the sample") , that is a follow up injection of mobile phase should never "dilute" any samples. I have no idea what you are referring to in this context. IN fact, for most types of analysis, a needle wash should not be needed unless the samples are rather 'sticky'.

You wrote: "The "binary" gradient is being used to mitigate the problems with frequent pump failure caused by air bubbles"; What gradient? It appears you are running isocratic with a RID (?). Pump cavitation is often caused by failure to maintain prime of the system. Common reasons for this include: a lack of continuous degassing of the mobile phase; sticking inlet or outlet valves; piston seal failure(s); clogged pickup frits; leaks or loose fittings. *If you are running a separate gradient step to wash the column down, what solvents are you using coupled to what time/flow program?

Thank you so much for your response. Regarding your questions:

1. The decision to use 150 uL was actually based on the 3x loop volume method. Unfortunately, the lab I've joined had misplaced all the papers containing the specifications of the HPLC including the loop volume. So, when I was developing the method I had to manually determine the loop volume by injecting liquid into the injector and measuring the subsequent volume of liquid drained/displaced from the injector. This volume was approx. 50 uL, hence the decision to use 150 uL injection (3x 50uL). It wasn't until much later that I found out that the loop volume was only 20 uL.

2. Yes, I did try a true blank for diagnosis. My approach was:
2.1 Inject a sample containing standards
2.2 Null injection which resulted in insignificant carryover, thus ruling out contamination in column.
2.3 True blank injection - where peaks for the standard compound from (2.1) were recorded with larger area than the sample/standard injection.

3. It's true that the sample injector hasn't been serviced in a very long time. I've requested the lab to schedule servicing soon. Since I'm a lowly Masters student, I'm not allowed to make any changes to the hardware.

4. We've only just installed a new column few weeks ago, so I don't think it's a major source of the issues.

5. Lactic acid is bit sticky/viscous.

6. Regarding the diluted peaks:
6.1 Since I usually analyze samples from different experimental conditions with high organic load (up to 25g/L), I observed carryover problems.
6.2 Following the null & true blank injections, I narrowed down the source of problem to the glass syringe and/or injector.
6.3 For the sample injectors (Rheodyne 7725i), I had to resort to "wash injections" using either true blank or distilled water to prevent compounds from one sample getting carried over to another.
6.4 Compared to older results for the same sample, the peak areas were significantly smaller for the sample prior to which I used wash injection. Since I use internal standards, I know that the peaks are in fact a lot smaller than they should be. But I don't know why.

7. I'm using the injector usage method I came across on this site and youtube.
7.1 Keep the injector in 'Load' position
7.2 Insert the syringe and inject the sample
7.3 Rotate the injector into the 'Inject' position to start the analysis but don't remove the syringe.
7.4 The injector and syringe stay in the same position as (7) until the end of the runtime.


If you don't mind I have a few stupid questions:
A. The peak areas will remain constant for all the sample injection volumes greater than the loop volume, is that correct? I.e In theory, a sample injection of 300 uL should give the same results as 150 uL & 20 uL (loop volume).

B. Is there any accepted method or protocol for cleaning/washing the rheodyne injectors after every (manual) sample injections?

Re-posting my previous questions to bump the threat:

1. Will the peak areas will remain constant for all the sample injection volumes greater than the loop volume? I.e In theory, a sample injection of 300 uL should give the same results as 150 uL & 20 uL (loop volume).

2. Are there any accepted methods or protocols for cleaning/washing the (rheodyne) injectors after every (manual) sample injections?

Hi Purplish,

You seem to have a doozy going on, and it sounds like your work is cut out for you at this new lab! If only they had maintained the instrument better!
As for your problem, I don't think I could be much help, however, I would recommend swapping out the parts you believe to be an issue, ie the syringe and injection port, and see if you get the same problems. Do you have an autosampler by any chance that you could check your manual injections against?

As for your questions, the reason the loop is filled to three times the volume is for chromatographic consistency. By filling the loop farther than the max volume, you assure that the loop is filled to it's maximum volume every time in order to get consistent injections. Excess volume will be sent to the waste. Additionally, while in the inject position, anything injected into the port will go directly to waste (if you want to clean it for example).

When it comes to cleaning the injector, it really depends on your analysis and the history of the injector. Seeing that you're at university, they may have less-rigorous standards when it comes to the system's upkeep and note-taking, so it may be difficult to track down what was injected on the column ( or at least very arduous ), but that will give you a better idea of what to use. If not, I'd probably start with 100% IPA if you're doing RP just because it's miscible in water, methanol and acetonitrile and is relatively non-polar so it can remove sticky stuff. When my autosampler's loop is too small, I frequently switch my samples to manual injections. I do prep work so my injector can look pretty disgusting at times (typically injecting 150 mg + / injection), and it frequently leaves a brown/black residue on my injection port AND my needle. The needle is also important to consider as it can be a main source of contamination-- especially if you're washing it with only water. I'd give the needle, pestel, etc. a nice long wash in some IPA as well.

Unfortunately, I don't think those will be the root of your problem, but hopefully it's somewhere to start and I hope I answered your questions!

As for your questions, the reason the loop is filled to three times the volume is for chromatographic consistency. By filling the loop farther than the max volume, you assure that the loop is filled to it's maximum volume every time in order to get consistent injections.

Just so I understand it correctly, an injection of 3 times the volume would give the same results as an injection of 5 times the volume. Is that correct?

Excess volume will be sent to the waste. Additionally, while in the inject position, anything injected into the port will go directly to waste (if you want to clean it for example).

Wouldn't the injected liquid go to the column when the injection port is in inject position?

Since this thread has come back to life, here's the explanation of 3 volumes:

(1) When you're loading the loop, you can push as much sample through the system as you want, because you're pushing sample into one end of the loop, and the other end is going to waste. When the valve changes to the 'inject' position, it cuts off the syringe at one end and the waste at the other, and puts just the loop into the flow-path, so one loop-volume gets injected, very precisely. That precision (repeatability) is the whole point of full-loop injection. So yes, 5 volumes should give the same answer as 3, because in the end, exactly 1 volume gets injected.

(2) Imagine the cross-section of the loop tubing: When you squirt the sample into the loop, the liquid nearest the edge of the tubing suffers a frictional drag as it slides along the tube. The liquid in the middle of the tube, furthest from the walls, suffers a lot less drag, so it moves faster. You don't get a flat interface between sample and solvent as you push the sample into the loop. Instead you get sample pushing forwards in the middle, while solvent lingers at the outside. This can be quite a serious linger! By pushing three loop volumes into the loop, you guarantee that even the bit right at the edge of the tubing will have been washed clean of solvent and now contain sample.

A few other thoughts: remember that the same applies to blank runs. They need three loop volumes of blank, too. And if you have bigger peaks on blanks than on samples in an isocratic run, do make sure that the peaks your looking at actually belong to the injection you think they do! If you have a ten minute run, and your sample takes 15 min to elute, then you'll have a peak at 5 minutes... but its size will correlate with the previous injection! Because that's where it came from...

But the likely sources of carry-over are (a) the syringe itself (get a new one, and use it only for blank solvent, never with samples or standards, until you have convinced yourself that you can get a blank that's really zero!). Or the other source, (b) is a badly-maintained valve, maybe one that's been turned while it had something horrible precipitated out in it, so it's got scratched, or one that's been taken to pieces by someone who didn't know what they were doing. I'd tend to put my bet on the syringe. Because it's an isocratic method, at least you can be sure the problem is something to do with injection, and not something nasty in the running buffer.

By the way, you shouldn't need to use a binary pump to pump one solvent because of bubbles. Those Shimadzu pumps have two heads in parallel, not in series, and they're rubbish at self-priming when dry. You'll already know the method for sucking from the purge waste tube with a syringe while purging to get rid of bubbles, but remember that with two parallel heads, you can end up with one of them bubble-free, while the other still has a bubble. If they're both bubble-free and the pressure is still not stable, I'd worry about check-valves.

... when you're in 'inject', the loop is in the flow-path from pump to column, and the syringe port is connected directly to the waste port. You'll never have the syringe connected directly to the column; the column is always connected, one way or the other, to the pump, and sitting at the full system pressure, while the syringe is always working at low pressure.